That's a good idea. Evidently an MOV can handle a lot of surge current, but it responds a little slowly. So the additional faster-acting protection catches the leading edge of the transients, for better protection.
For inrush current limiting, I have used 'Globar' resistors - negative temperature coefficient resistors that start out pretty high in value, and drop rapidly in resistance as they heat up. Placed in series with the AC power input, they greatly reduce the turn-on surge, without adding much resistance to the circuit in normal operation. However, they do not provide as much assistance as a series resistor if the AC power flickers, because the resistance of the Globar unit will be low as the power goes on and off.
These things operate HOT - so they need to be mounted appropriately, and wires and parts around them must be dressed away from them, so that they do not get scorched.
I don't see Globar in Google searches, but Digi-Key carries Cantherm inrush current limiters, which appear to use the same negative temperature coefficient concept.
http://www.cantherm.com/products/thermistors/cantherm_mf72.pdfMF72-003D9 - 4A, 3 ohms room temp, 0.12 ohms hot - Digi-Key #317-1156-ND
MF72-004D9 - 3A, 4 ohms room temp, 0.19 ohms hot - Digi-Key #317-1157-ND
MF72-005D9 - 3A, 5 ohms room temp, 0.21 ohms hot - Digi-Key #317-1158-ND
MF72-006D9 - 2A, 6 ohms room temp, 0.315 ohms hot - Digi-Key #317-1159-ND
MF72-008D9 - 2A, 8 ohms room temp, 0.40 ohms hot - Digi-Key #317-1160-ND
MF72-010D9 - 2A, 10 ohms room temp, 0.458 ohms hot - Digi-Key #317-1161-ND
http://www.digikey.comI have seen solid-state rectifiers with a turn-on delay. If this delay is based on the applied filament voltage, this would be very good for tube equipment, because it would allow the tube cathodes to be warmed before applying operating voltage.